Skip to content

Trend Prediction Service

Idea

The Trend Prediction Service predicts future values for time series using linear and nonlinear regression models. It is a forecasting framework, that has many useful applications in the area of Process & Condition Monitoring.

Typical use cases for the Trend Prediction Service are:

  • Predictive maintenance: Detect if a component's lifetime may be reached in the short-term future.
  • Monitoring of processes: Predict the duration of a process to prevent undesired states (e.g. waiting).
  • Seasonality and trend removal as preparation for other data analytics tasks.

Access

For accessing this service you need to have the respective roles listed in Analytics Services roles and scopes.

Basics

The Prediction Service is a data-driven approach that can be applied to univariate (single input variable) or multivariate (multiple input variables) time-series data. The predicted output is univariate and written into a single target variable.

The service provides the functionality required for estimating the relationships between the variables of a given time series in order to make predictions based on the trained model. These predictions can be used for reasoning about the process represented by the time series.

The trained models are based on a linear or polynomial regression.

Modes

Trend Prediction Service can be used in two modes:

Interactive Mode

This mode is intended to be used when user want to pass the training and prediction data in the request body. For example, when calling the API in a workflow from Visual Flow Creator. All the required configuration and data is provided in the request. The API calls are carried out synchronously and results are available immediately in the response.

The Model Management Service is used for model storage and automatically sets the expiration date of a model to 14 days. This parameter might be changed in the future.

Direct Interactive Mode

This mode is intended to be used when the user does not want to pass training and prediction data in the request body. In direct interactive mode, the Trend Prediction Service can communicate with IoT Time Series to fetch the training and prediction data. The user should provide the asset details and time range to the Trend Prediction Service API's instead of passing the training and prediction data in the request body. The API calls are carried out synchronously and results are available immediately in the response.

The Model Management Service is used for model storage and automatically sets the expiration date of a model to 14 days. This parameter might be changed in the future

Features

The Trend Prediction Service exposes its API for realizing the following tasks:

  • Train (fit) regression models using multivariate time-series data
  • Predict future values
  • Perform training and prediction using one request
  • Read and delete a regression models
  • Ability to communicate with IoT Time Series to fetch the training and prediction data for given asset details and time range

Limitations

  • In Interactive mode, the service can process a maximum of 20000 time series records.
  • In Direct Interactive mode, the service can process a maximum of 20000 time series records containing a maximum of 3 variables obtained from IoT Time Series service.

Example Scenario

An engineer monitoring the production line of a brewery wants to predict the expected energy consumption (y) for the next month. They assume a linear dependency between the energy consumption and time as well as the load, which are treated as independent variables.

The engineer collects time series data with the energy consumption and load of the production line, feeds the Trend Prediction Service API with this time series and evaluates the predicted future consumption.

Any questions left?

Ask the community


Except where otherwise noted, content on this site is licensed under the MindSphere Development License Agreement.